The purpose of the proposed study is to examine the effectiveness of two major approaches to teaching adolescents HIV prevention and sex education within the middle school environment: abstinence-only (i.e., abstinence until marriage) and safer- sex (i.e., focus on protection and ways to avoid risky sexual encounters). To do so, a comprehensive, multi-level approach will be utilized, including both classroom instruction and ongoing cohort-specific and school-wide programming intended to provide students with a consistent, sustained intervention message for two years. In addition, targeted programming to peer leaders, parents, teachers and other school officials (i.e., administrators, counselors, school nurses) will be included to enhance the individual-level intervention effects (DiClemente & Wingood, 2000). We propose to follow two cohorts of students (classes of 2009 and 2010; n=1800) from three ethnically diverse middle schools, enrolling each cohort into the study at the beginning of their 7"' grade school year (ages 12-13) and following them until the end of their 8" grade school year. Students in one school will receive the abstinence-until-marriage (using Operation Keepsake as the core curriculum), another will receive the safer-sex message (using Safer Choices), and the third school will serve as a non-intervention comparison group, receiving parallel programming in the area of general health promotion (using Get Connected). The specific aims of the study are (1) to determine both the short-term (6 month) and longer-term (12 mo., 18 mo.,) effects of the two different intervention approaches on individual behavior (i.e., sexual activity, use of protection, engagement in high risk situations), and the cognitive processes considered to mediate behavioral change (i.e., knowledge, beliefs, values, self- efficacy, preventive beliefs, and behavioral intentions); (2) to explore the impact of comprehensive and continuous programming on the sexual-related social norms as perceived by students, teachers, and other school officials; and (3) to explore the moderating effects of individual characteristics (e.g., gender, ethnicity, religion, prior sexual experience, school attendance/exposure to message) and teacher/administrator characteristics (e.g., belief in message, personal beliefs in regard to normative behavior of adolescents, involvement in and support of grant-sponsored programs) of the individual-level intervention effects on cognition and behavior. Multivariate analyses (path modeling, repeated measures component of HLM) will be utilized to compare the interventions to the comparison group and to each other, and to examine theoretical aspects of the behavioral change model.